
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import os


Online_Reviews_raw= pd.read_csv("Online_Reviews.csv")
Online_Reviews_raw.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 71044 entries, 0 to 71043
Data columns (total 25 columns):
 #   Column                Non-Null Count  Dtype  
---  ------                --------------  -----  
 0   id                    71044 non-null  object 
 1   brand                 71044 non-null  object 
 2   categories            71044 non-null  object 
 3   dateAdded             71044 non-null  object 
 4   dateUpdated           71044 non-null  object 
 5   ean                   39065 non-null  object 
 6   keys                  71044 non-null  object 
 7   manufacturer          70903 non-null  object 
 8   manufacturerNumber    70841 non-null  object 
 9   name                  71044 non-null  object 
 10  reviews.date          70977 non-null  object 
 11  reviews.dateAdded     71044 non-null  object 
 12  reviews.dateSeen      71044 non-null  object 
 13  reviews.didPurchase   32158 non-null  object 
 14  reviews.doRecommend   60429 non-null  object 
 15  reviews.id            32158 non-null  float64
 16  reviews.numHelpful    32508 non-null  float64
 17  reviews.rating        71044 non-null  int64  
 18  reviews.sourceURLs    71044 non-null  object 
 19  reviews.text          71008 non-null  object 
 20  reviews.title         70568 non-null  object 
 21  reviews.userCity      5410 non-null   object 
 22  reviews.userProvince  449 non-null    object 
 23  reviews.username      70948 non-null  object 
 24  upc                   71042 non-null  object 
dtypes: float64(2), int64(1), object(22)
memory usage: 13.6+ MB


Online_Reviews_raw.head(5)


Online_Reviews_raw.isnull().sum()

id                          0
brand                       0
categories                  0
dateAdded                   0
dateUpdated                 0
ean                     31979
keys                        0
manufacturer              141
manufacturerNumber        203
name                        0
reviews.date               67
reviews.dateAdded           0
reviews.dateSeen            0
reviews.didPurchase     38886
reviews.doRecommend     10615
reviews.id              38886
reviews.numHelpful      38536
reviews.rating              0
reviews.sourceURLs          0
reviews.text               36
reviews.title             476
reviews.userCity        65634
reviews.userProvince    70595
reviews.username           96
upc                         2
dtype: int64


Online_Reviews_raw.columns = Online_Reviews_raw.columns.str.replace(".", "_")

C:\Users\mviha\AppData\Local\Temp\ipykernel_21516\1538785711.py:1: FutureWarning: The default value of regex will change from True to False in a future version. In addition, single character regular expressions will *not* be treated as literal strings when regex=True.
  Online_Reviews_raw.columns = Online_Reviews_raw.columns.str.replace(".", "_")


Online_Reviews_raw.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 71044 entries, 0 to 71043
Data columns (total 25 columns):
 #   Column                Non-Null Count  Dtype  
---  ------                --------------  -----  
 0   id                    71044 non-null  object 
 1   brand                 71044 non-null  object 
 2   categories            71044 non-null  object 
 3   dateAdded             71044 non-null  object 
 4   dateUpdated           71044 non-null  object 
 5   ean                   39065 non-null  object 
 6   keys                  71044 non-null  object 
 7   manufacturer          70903 non-null  object 
 8   manufacturerNumber    70841 non-null  object 
 9   name                  71044 non-null  object 
 10  reviews_date          70977 non-null  object 
 11  reviews_dateAdded     71044 non-null  object 
 12  reviews_dateSeen      71044 non-null  object 
 13  reviews_didPurchase   32158 non-null  object 
 14  reviews_doRecommend   60429 non-null  object 
 15  reviews_id            32158 non-null  float64
 16  reviews_numHelpful    32508 non-null  float64
 17  reviews_rating        71044 non-null  int64  
 18  reviews_sourceURLs    71044 non-null  object 
 19  reviews_text          71008 non-null  object 
 20  reviews_title         70568 non-null  object 
 21  reviews_userCity      5410 non-null   object 
 22  reviews_userProvince  449 non-null    object 
 23  reviews_username      70948 non-null  object 
 24  upc                   71042 non-null  object 
dtypes: float64(2), int64(1), object(22)
memory usage: 13.6+ MB


Online_Reviews_raw['seller_name'] = Online_Reviews_raw['reviews_sourceURLs'].apply(lambda a: a.split('.com')[0].split('//')[-1] + '.com')
Online_Reviews_raw['seller_name']= Online_Reviews_raw['seller_name'].apply(lambda b: b.split('.')[-2])


Online_Reviews = Online_Reviews_raw[['id', 'brand', 'name','seller_name', 'reviews_rating', 'reviews_text']].copy()


Online_Reviews.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 71044 entries, 0 to 71043
Data columns (total 6 columns):
 #   Column          Non-Null Count  Dtype 
---  ------          --------------  ----- 
 0   id              71044 non-null  object
 1   brand           71044 non-null  object
 2   name            71044 non-null  object
 3   seller_name     71044 non-null  object
 4   reviews_rating  71044 non-null  int64 
 5   reviews_text    71008 non-null  object
dtypes: int64(1), object(5)
memory usage: 3.3+ MB


Online_Reviews.head(5)


Online_Reviews.isnull().sum()

id                 0
brand              0
name               0
seller_name        0
reviews_rating     0
reviews_text      36
dtype: int64


Online_Reviews.dropna(how='any', inplace=True)


Online_Reviews.isnull().sum()

id                0
brand             0
name              0
seller_name       0
reviews_rating    0
reviews_text      0
dtype: int64


import plotly.express as px


Online_Reviews_5rating= Online_Reviews[Online_Reviews['reviews_rating']==5]
Online_Reviews_1rating= Online_Reviews[Online_Reviews['reviews_rating']==1]

rating5_count =Online_Reviews_5rating["brand"].value_counts()
rating1_count =Online_Reviews_1rating["brand"].value_counts()


fig_1 = px.pie(rating5_count,values="brand",names=rating5_count.index)
fig_1.update_traces(textinfo="label+value+percent",textposition='inside')
fig_1.show()


fig_2 = px.pie(rating1_count,values="brand",names=rating1_count.index)
fig_2.update_traces(textinfo="label+value+percent",textposition='inside')
fig_2.show()


brand_summ = Online_Reviews.groupby('brand').agg({'reviews_rating': ['mean', 'count']}).reset_index()
brand_summ.columns = ['brand', 'mean_rating', 'total_reviews']
brand_summ = brand_summ[brand_summ['total_reviews'] > 200]
brand_summ


fig_3 = px.scatter(brand_summ,size='total_reviews', x='mean_rating', y='total_reviews', color='brand',hover_data=['brand'])
fig_3.show()


seller_summ = Online_Reviews.groupby('seller_name').agg({'reviews_rating': 'mean', 'reviews_text': 'count'}).reset_index()
seller_summ.columns = ['seller_name', 'mean_rating', 'total_reviews']
seller_summ=seller_summ[seller_summ["total_reviews"]>20]
fig_4 = px.pie(seller_summ, names='seller_name', values='mean_rating',hover_data=['mean_rating'])
fig_4.update_traces(textinfo='label+value+percent')
fig_4.show()


fig_4 = px.histogram(Online_Reviews, x='reviews_rating', nbins=15, color='seller_name')
fig_4.show()


rating_percentage = 100* Online_Reviews['reviews_rating'].value_counts()/len(Online_Reviews)
rating_percentage

5    65.508112
4    20.549797
3     6.150011
1     5.212089
2     2.579991
Name: reviews_rating, dtype: float64


number_of_words_per_review = Online_Reviews.reviews_text.str.split().str.len()
number_of_words_per_review.hist(bins=50, figsize=(15, 4))
plt.xlabel('Number of Words Per Review')
plt.ylabel('Freq')
plt.show()


import nltk
nltk.download('punkt') 
nltk.download('stopwords')

[nltk_data] Downloading package punkt to
[nltk_data]     C:\Users\mviha\AppData\Roaming\nltk_data...
[nltk_data]   Package punkt is already up-to-date!
[nltk_data] Downloading package stopwords to
[nltk_data]     C:\Users\mviha\AppData\Roaming\nltk_data...
[nltk_data]   Package stopwords is already up-to-date!

True


Reviews = Online_Reviews['reviews_text'].str.cat(sep=' ')


from nltk import word_tokenize


#function to split text into word
tokens = word_tokenize(Reviews)


frequency_dist = nltk.FreqDist(tokens)
top_50_tokens=sorted(frequency_dist,key=frequency_dist.__getitem__, reverse=True)[0:50]
# Create a bar chart
plt.figure(figsize=(12,6))
plt.bar(top_50_tokens, [frequency_dist[token] for token in top_50_tokens])
plt.xticks(rotation=90)
plt.xlabel('Token')
plt.ylabel('Frequency')
plt.title('Top 50 Most Frequent Tokens')
plt.show()


Online_Reviews['Sentiment'] = np.where(Online_Reviews.reviews_rating > 3,1,0)


Online_Reviews = Online_Reviews[Online_Reviews.reviews_rating != 3]  # Removing neutral reviews.


Online_Reviews


Online_Reviews['Sentiment'].value_counts()

1    61108
0     5533
Name: Sentiment, dtype: int64


Online_Reviews['lowercase_text_reviews'] = Online_Reviews['reviews_text'].str.lower()


token_list_before = [word_tokenize(i) for i in Online_Reviews['reviews_text']]
tokens = []
for sublist in token_list_before:
    for item in sublist:
        tokens.append(item)
print("Number of unique tokens then: ", len(set(tokens)))

Number of unique tokens then:  40322


token_list_before = [word_tokenize(i) for i in Online_Reviews['lowercase_text_reviews']]
tokens = []
for sublist in token_list_before:
    for item in sublist:
        tokens.append(item)
print("Number of unique tokens then: ", len(set(tokens)))

Number of unique tokens then:  32497


spl_chars = []
for review in Online_Reviews['lowercase_text_reviews']:
    chars = []
    for char in list(review):
        if not char.isalnum() and char != ' ':
            chars.append(char)
    spl_chars.append(chars)


Online_Reviews_old = Online_Reviews.copy()


print("Old Review: " + Online_Reviews_old['lowercase_text_reviews'][100])

Old Review: great product...no oily feeling...would recommend. this review was collected as part of a promotion.


 Online_Reviews['lowercase_text_reviews'] = Online_Reviews['lowercase_text_reviews'].str.replace(r'[^A-Za-z0-9 ]+', ' ')

C:\Users\mviha\AppData\Local\Temp\ipykernel_21516\1845484409.py:1: FutureWarning:

The default value of regex will change from True to False in a future version.


print("New Review : " + Online_Reviews['lowercase_text_reviews'][100])

New Review : great product no oily feeling would recommend  this review was collected as part of a promotion


nltk.download('stopwords')
from nltk.corpus import stopwords

[nltk_data] Downloading package stopwords to
[nltk_data]     C:\Users\mviha\AppData\Roaming\nltk_data...
[nltk_data]   Package stopwords is already up-to-date!


stop_words = stopwords.words('english')


def remove_stopwords(stop_words, sentence):
    return [word for word in nltk.word_tokenize(sentence) if word not in stop_words]

Online_Reviews['withoutstop_words'] = Online_Reviews['lowercase_text_reviews'].apply(lambda row: remove_stopwords(stop_words, row))
Online_Reviews[['lowercase_text_reviews','withoutstop_words']]


Online_Reviews


from nltk.stem import PorterStemmer
porter = PorterStemmer()


def stemming(sentence):
    tokenized_words=word_tokenize(sentence)
    stemmed_words=[]
    for word in tokenized_words:
        stemmed_words.append(porter.stem(word))
        stemmed_words.append(" ")
    return "".join(stemmed_words)

Online_Reviews['stemmed_words'] = Online_Reviews['withoutstop_words'].apply(lambda x: [stemming(y) for y in x])
Online_Reviews[['withoutstop_words','stemmed_words']]


from nltk.stem import LancasterStemmer
lancaster = LancasterStemmer()


from nltk.stem import WordNetLemmatizer 
lemmatizer = WordNetLemmatizer()


def lemmatization(sentence):
    tokenized_words=word_tokenize(sentence)
    lemmatized_words=[]
    for word in tokenized_words:
        lemmatized_words.append(lemmatizer.lemmatize(word, pos="v"))
        lemmatized_words.append(" ")
    return "".join(lemmatized_words)

Online_Reviews['lemmatized_words'] = Online_Reviews['withoutstop_words'].apply(lambda x: [lemmatization(y) for y in x])
Online_Reviews[['withoutstop_words','lemmatized_words']]


Online_Reviews


import sklearn


from sklearn.decomposition import LatentDirichletAllocation
from sklearn.feature_extraction.text import CountVectorizer


vectorizer = CountVectorizer(tokenizer=word_tokenize, stop_words=stopwords.words('english'))
td_matrix = vectorizer.fit_transform(Online_Reviews['lowercase_text_reviews'])

C:\Users\mviha\anakonda 3\lib\site-packages\sklearn\feature_extraction\text.py:528: UserWarning:

The parameter 'token_pattern' will not be used since 'tokenizer' is not None'

C:\Users\mviha\anakonda 3\lib\site-packages\sklearn\feature_extraction\text.py:409: UserWarning:

Your stop_words may be inconsistent with your preprocessing. Tokenizing the stop words generated tokens ["'d", "'ll", "'re", "'s", "'ve", 'could', 'might', 'must', "n't", 'need', 'sha', 'wo', 'would'] not in stop_words.


lda = LatentDirichletAllocation(n_components=5, random_state=0)
lda.fit(td_matrix)

LatentDirichletAllocation(n_components=5, random_state=0)

LatentDirichletAllocation(n_components=5, random_state=0)


# Print top words for each topic
def print_top_words(model, feature_names, n_top_words=15):
    for topic_idx, topic in enumerate(model.components_):
        print(f"Topic {topic_idx}:")
        print(" ".join([feature_names[i]
                        for i in topic.argsort()[:-n_top_words - 1:-1]]))
    print()


print_top_words(lda, vectorizer.get_feature_names_out(), n_top_words=15)

# Assign topics to reviews
topic_assignments = lda.transform(td_matrix)
Online_Reviews['topic'] = np.argmax(topic_assignments, axis=1)

Topic 0:
hair product review part conditioner promotion collected love like received soft shampoo great free used
Topic 1:
part review promotion collected use love clean wipes great product tide clorox easy clothes smell
Topic 2:
skin product review part promotion collected moisturizer face using use like olay feel really love
Topic 3:
use one product great mop like love bought would used get time spray easy well
Topic 4:
movie great good kids one family love watch funny loved movies really enjoyed first godzilla


vector = CountVectorizer()
Matrix = vector.fit_transform(Online_Reviews['lowercase_text_reviews'])
vector_df = pd.DataFrame(Matrix.toarray(), columns = vector.get_feature_names_out())
vector_df


bow_ct = CountVectorizer(tokenizer= word_tokenize,stop_words=stop_words,ngram_range=(1,3)) 
bow_data = bow_ct.fit_transform(Online_Reviews['lowercase_text_reviews'])
bow_data

C:\Users\mviha\anakonda 3\lib\site-packages\sklearn\feature_extraction\text.py:528: UserWarning:

The parameter 'token_pattern' will not be used since 'tokenizer' is not None'

C:\Users\mviha\anakonda 3\lib\site-packages\sklearn\feature_extraction\text.py:409: UserWarning:

Your stop_words may be inconsistent with your preprocessing. Tokenizing the stop words generated tokens ["'d", "'ll", "'re", "'s", "'ve", 'could', 'might', 'must', "n't", 'need', 'sha', 'wo', 'would'] not in stop_words.

<66641x1303171 sparse matrix of type '<class 'numpy.int64'>'
	with 3546540 stored elements in Compressed Sparse Row format>


from sklearn.model_selection import train_test_split


X_train_bow, X_test_bow, y_train_bow, y_test_bow = train_test_split(bow_data,Online_Reviews['Sentiment'],test_size = 0.25)


from sklearn.linear_model import LogisticRegression


logregg = LogisticRegression()
logregg.fit(X_train_bow, y_train_bow)

C:\Users\mviha\anakonda 3\lib\site-packages\sklearn\linear_model\_logistic.py:458: ConvergenceWarning:

lbfgs failed to converge (status=1):
STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.

Increase the number of iterations (max_iter) or scale the data as shown in:
    https://scikit-learn.org/stable/modules/preprocessing.html
Please also refer to the documentation for alternative solver options:
    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression

LogisticRegression()

LogisticRegression()


from sklearn.metrics import f1_score


logregg_pred = logregg.predict(X_test_bow)


print("f1 score for the model logistic_Regression : ", f1_score(y_test_bow, logregg_pred))

f1 score for the model logistic_Regression :  0.9806255458162176


from sklearn.ensemble import RandomForestClassifier


rf = RandomForestClassifier(random_state=42)
rf.fit(X_train_bow, y_train_bow)

RandomForestClassifier(random_state=42)

RandomForestClassifier(random_state=42)


rf_pred = rf.predict(X_test_bow)


print("f1 score: ", f1_score(y_test_bow,rf_pred))

f1 score:  0.9716795010500859


import gensim


# Load pre-trained GloVe model
model_path = os.path.join(os.getcwd(),'glove.42B.300d.txt')


model = gensim.models.KeyedVectors.load_word2vec_format(model_path, binary=False, no_header=True)


review_embed = [np.mean([model[i.lower()] for i in word_tokenize(review) if i.lower() in model], axis=0)
    if len([word for word in word_tokenize(review) if word.lower() in model]) > 0
    else np.zeros(model.vector_size)
    for review in Online_Reviews['lowercase_text_reviews']]
embed_data = pd.DataFrame(review_embed)


embed_data = embed_data.fillna(0)
embed_data


X_train_embed, X_test_embed, y_train_embed, y_test_embed = train_test_split(embed_data,Online_Reviews['Sentiment'],test_size = 0.20)


logregg_2 = LogisticRegression()


logregg_2.fit(X_train_embed, y_train_embed)
WE_logregg_pred = logregg_2.predict(X_test_embed)

print("f1 score for we: ",f1_score(y_test_embed, WE_logregg_pred))

f1 score for we:  0.969496981891348

C:\Users\mviha\anakonda 3\lib\site-packages\sklearn\linear_model\_logistic.py:458: ConvergenceWarning:

lbfgs failed to converge (status=1):
STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.

Increase the number of iterations (max_iter) or scale the data as shown in:
    https://scikit-learn.org/stable/modules/preprocessing.html
Please also refer to the documentation for alternative solver options:
    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression


import joblib

# save trained logistic regression model
joblib.dump(logregg, 'logregg.joblib')

# save CountVectorizer object
joblib.dump(bow_ct, 'bow_ct.joblib')

['bow_ct.joblib']


# function to preprocess user input
def preprocess_input(input_str):
    lowercase_str = input_str.lower() # convert to lowercase
    tokenized_str = word_tokenize(lowercase_str) # tokenize words
    return ' '.join(tokenized_str) # join tokenized words with spaces

# take user input
user_input = input("Enter your review: ")

# preprocess user input
preprocessed_input = preprocess_input(user_input)

# create bag-of-words representation of user input
input_bow = bow_ct.transform([preprocessed_input])

# predict sentiment rating
sentiment_prediction = logregg.predict(input_bow)

# print predicted sentiment rating
print("Predicted sentiment rating: ", sentiment_prediction[0])

Enter your review: Hi I brought eight desk chairs from this company in July 2022. In less than six months, the casters broke on fours of the chairs with light use. I contacted the company and was told it was passed the warranty period! Avoid this product and this company.
Predicted sentiment rating:  0

	id	brand	categories	dateAdded	dateUpdated	ean	keys	manufacturer	manufacturerNumber	name	...	reviews.id	reviews.numHelpful	reviews.rating	reviews.sourceURLs	reviews.text	reviews.title	reviews.userCity	reviews.userProvince	reviews.username	upc
0	AV13O1A8GV-KLJ3akUyj	Universal Music	Movies, Music & Books,Music,R&b,Movies & TV,Mo...	2017-07-25T00:52:42Z	2018-02-05T08:36:58Z	6.02537E+11	602537205981,universalmusic/14331328,universal...	Universal Music Group / Cash Money	14331328	Pink Friday: Roman Reloaded Re-Up (w/dvd)	...	NaN	0.0	5	https://redsky.target.com/groot-domain-api/v1/...	i love this album. it's very good. more to the...	Just Awesome	Los Angeles	NaN	Joshua	6.02537E+11
1	AV14LG0R-jtxr-f38QfS	Lundberg	Food,Packaged Foods,Snacks,Crackers,Snacks, Co...	2017-07-25T05:16:03Z	2018-02-05T11:27:45Z	73416000391	lundbergorganiccinnamontoastricecakes/b000fvzw...	Lundberg	574764	Lundberg Organic Cinnamon Toast Rice Cakes	...	100209113.0	NaN	5	https://www.walmart.com/reviews/product/29775278	Good flavor. This review was collected as part...	Good	NaN	NaN	Dorothy W	73416000391
2	AV14LG0R-jtxr-f38QfS	Lundberg	Food,Packaged Foods,Snacks,Crackers,Snacks, Co...	2017-07-25T05:16:03Z	2018-02-05T11:27:45Z	73416000391	lundbergorganiccinnamontoastricecakes/b000fvzw...	Lundberg	574764	Lundberg Organic Cinnamon Toast Rice Cakes	...	100209113.0	NaN	5	https://www.walmart.com/reviews/product/29775278	Good flavor.	Good	NaN	NaN	Dorothy W	73416000391
3	AV16khLE-jtxr-f38VFn	K-Y	Personal Care,Medicine Cabinet,Lubricant/Sperm...	2017-07-25T16:26:19Z	2018-02-05T11:25:51Z	67981934427	kylovesensualitypleasuregel/b00u2whx8s,0679819...	K-Y	67981934427	K-Y Love Sensuality Pleasure Gel	...	113026909.0	NaN	1	https://www.walmart.com/reviews/product/43383370	I read through the reviews on here before look...	Disappointed	NaN	NaN	Rebecca	67981934427
4	AV16khLE-jtxr-f38VFn	K-Y	Personal Care,Medicine Cabinet,Lubricant/Sperm...	2017-07-25T16:26:19Z	2018-02-05T11:25:51Z	67981934427	kylovesensualitypleasuregel/b00u2whx8s,0679819...	K-Y	67981934427	K-Y Love Sensuality Pleasure Gel	...	171267657.0	NaN	1	https://www.walmart.com/reviews/product/43383370	My husband bought this gel for us. The gel cau...	Irritation	NaN	NaN	Walker557	67981934427

	id	brand	name	seller_name	reviews_rating	reviews_text
0	AV13O1A8GV-KLJ3akUyj	Universal Music	Pink Friday: Roman Reloaded Re-Up (w/dvd)	target	5	i love this album. it's very good. more to the...
1	AV14LG0R-jtxr-f38QfS	Lundberg	Lundberg Organic Cinnamon Toast Rice Cakes	walmart	5	Good flavor. This review was collected as part...
2	AV14LG0R-jtxr-f38QfS	Lundberg	Lundberg Organic Cinnamon Toast Rice Cakes	walmart	5	Good flavor.
3	AV16khLE-jtxr-f38VFn	K-Y	K-Y Love Sensuality Pleasure Gel	walmart	1	I read through the reviews on here before look...
4	AV16khLE-jtxr-f38VFn	K-Y	K-Y Love Sensuality Pleasure Gel	walmart	1	My husband bought this gel for us. The gel cau...

	brand	mean_rating	total_reviews
22	Aveeno	3.807692	442
23	Avery	4.886076	316
53	Burt's Bees	4.630052	965
78	Clear Scalp & Hair Therapy	2.816976	377
80	Clorox	4.825421	10700
96	Disney	4.722914	3692
105	FOX	4.107603	4498
113	Foodsaver	4.310051	587
122	Head & Shoulders	4.307965	565
134	Honeywell	4.290323	310
135	Hoover	4.397849	372
137	Hormel	4.183267	251
163	Just For Men	4.189904	416
171	Kiss	4.411032	562
176	L'Oreal Paris	4.352218	2683
177	L'oreal Paris	4.495629	2288
184	Lemi Shine	4.794326	423
185	Lionsgate	4.740192	1351
190	Lundberg	4.772300	426
193	Lysol	4.369048	336
200	Maybelline	3.613208	742
226	Nexxus	1.549806	773
228	Nivea	4.576577	222
241	Olay	4.193567	3420
246	Opti-Free	4.251121	223
255	Pantene	4.239695	655
256	Paramount Pictures	4.579268	328
259	Pendaflex	4.165761	368
275	Quilted Northern	4.826268	1439
282	Rubbermaid	2.801444	1939
291	Samsung	3.759146	328
313	Sony	4.159459	370
314	Sony Music	4.685714	245
315	Sony Pictures	4.357242	2161
325	Storkcraft	4.236842	266
327	Suave	4.134529	223
329	Summit Entertainment	4.578563	821
337	Target.com Use Only	4.522659	331
343	Tide	4.168091	5384
348	Tostitos	4.623693	287
360	Universal Home Video	4.663807	6178
364	Vaseline	4.744186	258
370	Warner Bros.	4.538542	1920
371	Warner Home Video	4.349550	3330
373	Warner Music Group	4.692683	205
382	Windex	4.129310	348

	id	brand	name	seller_name	reviews_rating	reviews_text	Sentiment
0	AV13O1A8GV-KLJ3akUyj	Universal Music	Pink Friday: Roman Reloaded Re-Up (w/dvd)	target	5	i love this album. it's very good. more to the...	1
1	AV14LG0R-jtxr-f38QfS	Lundberg	Lundberg Organic Cinnamon Toast Rice Cakes	walmart	5	Good flavor. This review was collected as part...	1
2	AV14LG0R-jtxr-f38QfS	Lundberg	Lundberg Organic Cinnamon Toast Rice Cakes	walmart	5	Good flavor.	1
3	AV16khLE-jtxr-f38VFn	K-Y	K-Y Love Sensuality Pleasure Gel	walmart	1	I read through the reviews on here before look...	0
4	AV16khLE-jtxr-f38VFn	K-Y	K-Y Love Sensuality Pleasure Gel	walmart	1	My husband bought this gel for us. The gel cau...	0
...	...	...	...	...	...	...	...
71039	AVzRGDlw-jtxr-f3yfFH	Opti-Free	Opti-Free Puremoist Disinfecting Solution 10fl Oz	walmart	5	I'm dealing with dry eyes and very sensitive t...	1
71040	AVzRGDlw-jtxr-f3yfFH	Opti-Free	Opti-Free Puremoist Disinfecting Solution 10fl Oz	walmart	5	Great stuff! Helps keep my contacts moist all ...	1
71041	AVzRGDlw-jtxr-f3yfFH	Opti-Free	Opti-Free Puremoist Disinfecting Solution 10fl Oz	walmart	5	Right price	1
71042	AVzvTLfivKc47QAVfPme	Bebe Au Lait	Bebe Au Lait174 Premium Muslin Nursing Scarf -...	target	5	I was so pleased when this scarf arrived for m...	1
71043	AVzvWCVKglJLPUi8Gzji	UMGD	Justin Bieber - Under The Mistletoe (vinyl)	target	5	Excellent album, makes the holidays better, gi...	1

	lowercase_text_reviews	withoutstop_words
0	i love this album it s very good more to the...	[love, album, good, hip, hop, side, current, p...
1	good flavor this review was collected as part...	[good, flavor, review, collected, part, promot...
2	good flavor	[good, flavor]
3	i read through the reviews on here before look...	[read, reviews, looking, buying, one, couples,...
4	my husband bought this gel for us the gel cau...	[husband, bought, gel, us, gel, caused, irrita...
...	...	...
71039	i m dealing with dry eyes and very sensitive t...	[dealing, dry, eyes, sensitive, products, eye,...
71040	great stuff helps keep my contacts moist all ...	[great, stuff, helps, keep, contacts, moist, day]
71041	right price	[right, price]
71042	i was so pleased when this scarf arrived for m...	[pleased, scarf, arrived, multiple, reasons, p...
71043	excellent album makes the holidays better gi...	[excellent, album, makes, holidays, better, gi...

Sentimental Analysis On Online Reviews¶

Importing packages¶

Importing Data¶

Visualizations¶

Counting the total number of each brand's products¶

Creating a pie graph showing the brand counts.¶

Distribution of 5-star ratings by brand¶

Distribution of 1-star ratings by brand¶

Determining the average rating for each brand which has atleast 200 reviews¶

Determining the average rating for each seller which has atleast 20 reviews¶

Getting the number of words by splitting them by a space¶

NLP- Data Preprocessing¶

Identifying the most commonly used words in the reviews¶

Converting text reviews to lowercase:¶

Number of tokens before converting reviews to lowercase¶

Number of tokens after converting reviews to lowercase¶

Removing special characters¶

Stemming¶

lemmatization¶

Topic Modelling¶

Bag of Words Model¶

converting reviews to a matrix of token counts¶

Logistic regression¶

RandomForestClassifier¶

Word Embeddings Model¶

source : https://nlp.stanford.edu/projects/glove/¶

Creating Web page using streamlit¶

	withoutstop_words	stemmed_words
0	[love, album, good, hip, hop, side, current, p...	[love , album , good , hip , hop , side , curr...
1	[good, flavor, review, collected, part, promot...	[good , flavor , review , collect , part , pro...
2	[good, flavor]	[good , flavor ]
3	[read, reviews, looking, buying, one, couples,...	[read , review , look , buy , one , coupl , lu...
4	[husband, bought, gel, us, gel, caused, irrita...	[husband , bought , gel , us , gel , caus , ir...
...	...	...
71039	[dealing, dry, eyes, sensitive, products, eye,...	[deal , dri , eye , sensit , product , eye , c...
71040	[great, stuff, helps, keep, contacts, moist, day]	[great , stuff , help , keep , contact , moist...
71041	[right, price]	[right , price ]
71042	[pleased, scarf, arrived, multiple, reasons, p...	[pleas , scarf , arriv , multipl , reason , pr...
71043	[excellent, album, makes, holidays, better, gi...	[excel , album , make , holiday , better , giv...

	00	000	0000	000x	007	02	04	05	06	07	...	zoom	zootopia	zorbas	zow	zre	zucchetta	zucchini	zumba	zyliss	zyrtec
0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
2	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
3	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
4	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
66636	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
66637	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
66638	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
66639	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
66640	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0

	0	1	2	3	4	5	6	7	8	9	...	290	291	292	293	294	295	296	297	298	299
0	0.084472	0.006689	-0.030360	-0.023078	0.114714	-0.030096	-3.543177	0.063921	-0.044364	-0.466326	...	0.015356	-0.136490	-0.167926	0.014822	0.042434	0.075543	-0.092296	-0.087358	0.075571	0.083943
1	0.006620	0.155284	0.050900	-0.143476	0.361350	-0.187796	-3.327837	0.379316	-0.183479	-0.332947	...	-0.109184	-0.053782	-0.073997	0.013061	0.110625	0.008677	-0.087000	-0.083884	-0.118944	0.056691
2	-0.117517	0.145745	0.361120	-0.177808	0.653885	-0.314230	-3.131000	0.128090	-0.222540	-0.922825	...	-0.434730	-0.379955	0.225732	0.425675	0.543885	0.208977	-0.414850	-0.402450	-0.019839	0.105304
3	-0.055465	-0.015274	-0.081501	-0.127692	0.108524	-0.040868	-3.521162	0.251690	0.032892	-0.490006	...	-0.018850	-0.125914	-0.106902	-0.033968	0.097602	0.006678	-0.001957	-0.092296	0.017657	-0.048639
4	-0.079086	-0.031289	-0.098909	-0.134346	0.193464	-0.155696	-3.289908	0.282952	-0.107354	-0.637257	...	0.000952	-0.118683	-0.126174	-0.102727	0.139136	0.034631	-0.017657	-0.239479	-0.028623	0.041141
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
66636	-0.035993	0.005894	-0.070501	-0.135417	0.180412	0.020753	-3.663653	0.273143	0.059003	-0.470564	...	-0.037995	-0.158815	-0.040988	-0.121511	0.066875	-0.009848	0.009759	-0.095551	0.076208	0.022629
66637	0.015176	0.059635	-0.078321	-0.338198	0.166626	-0.068326	-3.375389	-0.030400	-0.026978	-0.468171	...	-0.060907	0.007270	-0.116024	-0.202493	0.038875	0.063515	-0.074470	0.007896	-0.072332	0.089074
66638	0.064952	0.059367	-0.109067	-0.259950	0.404300	-0.214533	-3.214650	0.291120	-0.172155	-0.558130	...	-0.340420	0.322485	-0.065600	0.077256	0.093956	0.029065	-0.156420	-0.258429	0.341340	0.001826
66639	0.017964	-0.065302	-0.113183	-0.189686	0.172825	-0.046762	-3.459496	0.154399	-0.006493	-0.489411	...	-0.012438	-0.146012	-0.105633	-0.039317	0.034064	0.088318	-0.006881	-0.161544	0.087350	-0.003968
66640	0.075640	0.000355	-0.123217	0.083338	0.180645	-0.059262	-3.166650	-0.128553	-0.013581	-0.525699	...	0.024206	-0.055767	-0.127152	-0.131488	0.019077	0.055419	0.080246	-0.018043	0.055248	-0.066320